Methods are herein provided for decision support in diagnosis of a disease in a subject, and for extracting features from a multi-slice data set. Systems for computer-aided diagnosis are provided. The systems take as input a plurality of medical data and produces as output a diagnosis based upon this data. The inputs may consist of a combination of image data and clinical data. Diagnosis is performed through feature selection and the use of one or more classifier algorithms.
Legal claims defining the scope of protection, as filed with the USPTO.
1. A method of extracting features from a multi-slice data set, the method comprising: representing a spatial distribution of an object mathematically; representing a shape of the object mathematically; determining contour and texture of the object; identifying a border pixel of the object and estimating a derivative; analyzing the derivative as a function of position along the contour; identifying automatically the presence of dark regions or bright regions within the object; and approximating texture of an image in a surrounding region of the object; wherein the features are calculated for the group comprising each slice of the multi-slice data set, a maximum intensity projection taken at an arbitrary angle, a minimum intensity projection taken at an arbitrary angle, and a digitally reconstructed radiograph taken at an arbitrary angle through one or more slices of the image.
2. The method according to claim 1 , further comprising selecting an individual slice from the multi-slice data set for analysis by a manual selection by a user or by an automatic selection of a largest slice.
3. The method according to claim 1 , wherein the features calculated for each slice of the multi-slice data set are combined by a method selected from the group consisting of: calculating a weighted average in which weights are proportional to a number of pixels on each slice; finding a maximum value across multiple slices of the multi-slice data set; and finding a minimum value across the multiple slices of the multi-slice data set.
4. The method according to claim 1 , wherein the features are calculated in each of a plurality of dimensions.
5. The method according to claim 4 , wherein the plurality of dimensions is at least one selected from the group consisting of 2 dimensions, 2.5 dimensions, and 3 dimension.
6. The method according to claim 1 , wherein the shape of the object is described by at least one of the group consisting of: distribution of coefficients after a Fourier transform of border pixel positions; mathematical moments of a segmented object that are invariant to translation, rotation, and scaling; mathematical moments of a grayscale distribution of image pixels; fractal dimension; and a chain code.
7. The method according to claim 1 , wherein the texture of the object is described by at least one of the group consisting of: fractal dimension; energy, entropy, maximum probability, inertia, inverse difference and correlation based on a gray-level co-occurrence matrix; and coarseness, contrast, busyness, complexity and strength based on a neighborhood gray-tone difference matrix.
8. The method according to claim 1 , wherein the surrounding region is described by at least one of the group consisting of: a derivative of image intensity along a direction orthogonal to a local contour; a derivative of the image intensity along the direction orthogonal to the local contour and moments of a power spectrum; and an estimate of variance of the image intensity along the direction orthogonal to the local contour.
9. The method according to claim 1 , wherein the presence of dark regions and bright regions within the object is described by the intensity or size of clusters of contiguous pixels above or below a given threshold.
10. A system for extracting features from a multi-slice data set, the system comprising: a processor; and a memory storing instructions, which, when executed by the processor, cause the processor to: represent a spatial distribution of an object mathematically; 0067 represent a shape of the object mathematically; determine contour and texture of the object; identify a border pixel of the object and estimate a derivative; analyze the derivative as a function of position along the contour; identify automatically the presence of dark regions or bright regions within the object; and approximate texture of an image in a surrounding region of the object; wherein the features are calculated for the group comprising each slice of the multi-slice data set, a maximum intensity projection taken at an arbitrary angle, a minimum intensity projection taken at an arbitrary angle; and a digitally reconstructed radiograph taken at an arbitrary angle through one or more slices of the image.
11. The system according to claim 10 , wherein the memory further stores instructions, which, when executed by the processor, cause the processor to select an individual slice from the multi-slice data set for analysis by a manual selection by a user or by an automatic selection of a largest slice.
12. The system according to claim 10 , wherein the features calculated for each slice of the multi-slice data set are combined by one of the group consisting of: calculating a weighted average in which weights are proportional to a number of pixels on each slice; finding a maximum value across multiple slices of the multi-slice data set; and finding a minimum value across the multiple slices of the multi-slice data set.
13. The system according to claim 10 , wherein the features are calculated in each of a plurality of dimensions.
14. The system according to claim 13 , wherein the plurality of dimensions is at least one selected from the group consisting of 2 dimensions, 2.5 dimensions, and 3 dimension.
15. The system according to claim 10 , wherein the shape of the object is described by at least one of the group consisting of: distribution of coefficients after a Fourier transform of border pixel positions; mathematical moments of a segmented object that are invariant to translation, rotation, and scaling; mathematical moments of a grayscale distribution of image pixels; fractal dimension; and a chain code.
16. The system according to claim 10 , wherein the texture of the object is described by at least one of the group consisting of: fractal dimension; energy, entropy, maximum probability, inertia, inverse difference and correlation based on a gray-level co-occurrence matrix; and coarseness, contrast, busyness, complexity and strength based on a neighborhood gray-tone difference matrix.
17. The system according to claim 10 , wherein the surrounding region is described by at least one of the group consisting of: a derivative of image intensity along a direction orthogonal to a local contour; a derivative of the image intensity along the direction orthogonal to the local contour and moments of a power spectrum; and an estimate of variance of the image intensity along the direction orthogonal to the local contour.
18. The system according to claim 10 , wherein the presence of dark regions and bright regions within the object is described by the intensity or size of clusters of contiguous pixels above or below a given threshold.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
October 5, 2018
May 11, 2021
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.